Recently put together a new system with the following spec:
Athlon x2 4050e Processor 2.1ghz
Kingston 2x1GB DDR2 800MHz/PC2-6400 Memory Non-ecc CL5 Unbuffered 1.8V
Gigabyte GA-MA69VM-S2 Motherboard
Antec 500w PSU
Optiarc AD-5200A-0B 20X DVD±RW/DL Int IDE Bare Black Drive - OEM
Seagate ST3250310AS 250GB Hard Drive SATAII 7200rpm 8MB Cache - OEM
Went together fine and set it up with XP SP3, installed software and was all ok until it restarted itself. Checked event viewer and got the 1003 error.
Couldnt get any other info really from it, i changed the recovery settings so i got a BSOD rather than a reboot and then began testing. The hdd passed seagate's software testing, ran memtest for 3 passes all ok, ran some stress test and cpu on its own passed no problem. It did fail on the cpu and memory blend test? i forget the name of the tester and dont have the link :( So that could shed some light maybe? just need to find the test i had.....
I had a spare new PSU so swapped that in and its still BSOD after that so i rule out the PSU. Its not a massive spec so isnt really using much power.
Its happened three times in last 4 days now. i have attached the minidumps to see if that offers anything useful.
3 different faults for those minidumps which is strange the first usually points to malware/ spyware could be a trojan so download hijack this and some anti spyware and run full scans, post your hi jack thsi log back here and take it from there
Yes, the three different minidump results could be due to RAM or virus/Malware. Although you've already checked memory, does it consist of more than one RAM stick? If yes, you could try removing each RAM in turn, & see if that resolves the reboots & BSOD.
Memory has the 2x1gb sticks - ran some tests on the PC with 1 stick in and couldnt get BSOD, ran with other stick in and also couldnt get BSOD (burning cd, defrag, pc mark 05) put both in and again couldnt get it to BSOD. Problem is crashes before been quite random so unsure if re-seating ram has solved issue or if i havent been able to replicate problem.
Ok. Well experience shows that a RAM test such as http://www.memtest.org/ needs at least 3 passes to get a reasonable result, and even then there's no guarantee.
The suggestion of running with one stick at a time was to prove beyond any reasonable doubt. However it does look as though you may have to run for several hours(maybe overnight) with each RAM in turn, to definitely confirm.
Meanwhile it's still worth running those two earlier scanners when appropriate, and let slam69 see the HijackThis logfile.
Logfile of Trend Micro HijackThis v2.0.2
Scan saved at 10:15:41, on 22/07/2008
Platform: Windows XP SP3 (WinNT 5.01.2600)
MSIE: Internet Explorer v7.00 (7.00.6000.16674)
Boot mode: Normal
For the HJT logfile i absolutely agree with slam69. However, if the RAMs eventually prove to be ok then one logfile entry that HijackThis sees as slightly 'suspicious', is this next one.
However i'm pretty sure slam69 would have seen it, & it's almost certainly safe! The nearest recognition i can find is for Cardiff University.
*If* you don't recognise it, let us know & we'll try & trace it further>
Therefore, after your RAM tests, you may like to try the Kaspersky free online virus scanner, which is a good way to find out if you have any viruses or spyware without having to uninstall your existing antivirus software> http://www.kaspersky.co.uk/virusscanner
yeah saw that one but saw the asker is based in UK and teh ac segment pointed me towards a uni, i went with university of wales but close enough :O) havent ever seen an errant proxy server virus setup utilising an ac address so discounted ;O)
Hi just to clarify i work for UWIC (University of Wales Instittute Cardiff) and PC was connected to the internal network here a while back and so auto detected the proxy server.
So thats perfectly fine. Going to run a memtest overnight on offending PC and check results of that.
@ jimmyca69 ... ok, good, although i'd still consider running Kaspersky free, tomorrow.
Then, in the unlikely case that it's still unresolved, we could take a look at another group of minidumps.
Just a quick note, have PC with me and running memtest as we speak.
Notice in memtest it says the ram is running @ 421mhz meaning DDR842mhz in total. Now is this memtest reading it incorrectly? just wondering if technically its being overclocked by the bios? and hence causing problems. Anyway assume this would cause a failure in memtest so will leave it running for a couple of hours at least.
Have tried to analyze your two minidumps but i'm getting "corrupt or in a format not understood by the debugger" errors'. Could be this machine so investigating further.
Incidently, the suggestion of running with one stick at a time may still be an option, even though Memtest reports all ok.
According to bios and also the easytune software that comes with the mobo the cpu temp rarely goes above 30 and the system never above 35. Even when i was testing cpu the cpu temp never got above 40 on full load as i was monitoring it.
Any recommended cpu testers i can try to confirm cpu ok?
I downloaded that smartfan but it doesnt seem to pick up ay temps other than that of HDD which is 38 degrees.
I ran the orthos test again, the blend test this time ran 20mins without error and also ran hotcpu and it ran for 30 mins all ok.
I am in process of re-installing all the latest drivers from the mobo website to make sure they are ok. Apparently XP SP3 has habit of replacing some drivers and causing some BSOD on AMD chipsets so im trying that out now. As thinking about it these problems only occurred since SP3 was installed but then again i did install it soon after the OS went on.
well you could be onto something then, however the internal diagnostics is a set of tools set up by the manufacturer to test teh onboard ardware, if you log onto teh manufacturers website you shoudl be able to find them usually a disk you plug in and then off you go, usually included is a fan test piece of software which puts teh internal fan through a series of high rotations to see if it is functioning correctly, often when overheating is occuring it is due to teh fan not reaching teh high stages of its revolutions
ah right ok - with this board i have easytune that monitors fan and cpu.
Currently whilst running orthos cpu blend test the system temp after 3 mins is 39, cpu is 39 and cpu fan running at 2295rpm.
Whilst watching it stopped after 3 mins citing a hardware failure - temp never got above 39 and once stopped cpu dropped back to 30 eventually. Its not crashing the system just stopping the test, but not sure what its saying is wrong. Error message i get is :
Using CPU #1
Beginning a continuous self-test to check your computer.
Press Stop to end this test.
Test 1, 4000 Lucas-Lehmer iterations of M19922945 using 1024K FFT length.
FATAL ERROR: Rounding was 0.4951171875, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Torture Test ran 3 minutes 39 seconds - 1 errors, 0 warnings.
Execution halted.
The stress.txt file is supposed to be in the same folder as the .exe file, but reports seem to indicate that a number of people are unable to locate this txt file!
Update: Ran memtest overnight and got 23 passes so all ok.
Just to check the settings i mentioned previously about the ram - is it ok that its running @ DDR842 (421mhz each stick) i assume given the passes that yes it is ok?
I can check the volts via the Easytune software and when running the following range is shown
Vcore = 1.170-1.190
+3.3v = 3.340
+12v = 12.170-12.230
CPU Vid = 1.200
From your overnight RAM test results then yes, i too assume that all is ok.
All along i've had a slight suspicion about the cooling, and i agree with slam69's thoughts. You may even want to consider exchanging that fan (if convenient at this time)before going any further.
From reading around the subject of overclocking, and ~if~ your problem remained unresolved, you could take a look at the Vcore voltage(1.170-1.190). Suggest a slight tweak to 1.20 or even 1.21, and if unsuccessful a decrease tweak to 1.15, to check for any improvements.
thought i had got somewhere with us - all the tests done yesterday and this morning and no BSOD. Began everday testing by burning CD, downloading some game demo's and installing them just to push machine. Was all going so well and then BSOD :( Attached is the dump text but doubt it will reveal anything.
I couldnt find gigabyte's internal diagnostics, there is the easytune 6 software that gives cpu temp, voltages, fan speed etc but nothing specific for fan that i can find.
Also found the BIOS for this board to be most unhelpful, no overclocking options to change ram timings, voltages etc etc which i find odd. minidump240708.txt
quick update - was able to use the ctrl F1 function to get the advanced bios settings for the mobo but still cant find options to increase the vcore voltage.
I did however change the bios settings so that the cpu fan was not being controlled by the bios and so it was on full at all times and now constantly runs at 3500rpm. Did a search for cpu fan testing software and its throwing up nothing.
ok not ideal but keep the fan running at the top speed and wait 24 hours conduction usual activites as before and see if you get a BSOD if none then we can start to assume its yoru fan, and then we havea couple of options, either exchanging teh current on board one or fitting an additional, but lets see what happens
Wanted to rule out software issue so re-installed OS and literally after only 15mins of installing board drivers it crashed once again with same error.
Tearing my hair out now - beginning to suspect faulty cpu or motherboard. Problem is the company I bought from will test any returned products. Unsure how they will replicate problem before saying thats its not faulty and charging me for it.
if it is CPU then why didnt throw up errors on hotcpu or orthos tests?
If it is mobo - then why did memtest pass, cpu tests pass etc as all these running through the motherboard.
Any other suggestions of what maybe to try to root out the cause of this? Tempted to just buy a new motherboard and if that works then send old one back as only 7 days old and have 28 days to send stuff back.
this is becoming frustrating, has to be hardware as OS reinstalled iw ould return the unit as a whole and advise them of teh faults take in teh minidumps with you as proof of your issue and that you have tried to pinpoint it yourself but have exhausted all avenues, if they test it they shoudl have no issue recreating the fault, even sit in on the test if you can spare the time
id love to be able to give you the issue but from over the minitor assistance im running out of ideas
the variation in files within teh minidum,ps to me indicates that it is justa load on teh machien causing it to blue creen causing the variation in causes due to whatever activity is happening at the time
@ jimmyca69 ... the last dump was a .txt file.
Also, no problem, but you don't appear to have the correct symbols for WinDbg.
For later, if you wish to analyse your own dump file>
"How to read the small memory dump files that Windows creates for debugging": http://support.microsoft.com/kb/315263
The dumps are normally located in c:\windows\minidump\
or %systemroot%\minidump\
@ slam69 Agreed, nothing 'visible' in HijackThis logfile, but although perhaps unlikely, there could still be other nasties present, eg rootkit. This would not necessarily show up in the HJT log.
The Kaspersky free online virus scanner is not an 'absolute guarantee' of a clean system, but it is a pretty good way to find out if you have any viruses or spyware without having to uninstall your existing antivirus software. You may want to try this one, 'when convenient'> http://www.kaspersky.co.uk/virusscanner
We've more powerful tools/scanners *if* a rootkit was suspected.
Have run the kaspersky scan and its clean (it should be as hasnt even been on web or network yet!)
Unfortunately i only saved the txt file versions of the dmp files to my pen drive and now after re-installing OS i lost the originals. I do have a couple that occured since the rebuild which i will post up as soon as seagate tools finish testing drive.
I'd love it to be something like a rootkit but i just cant see it. NAV was installed first time before joining network and had latest virus defs all setup. Plus i was behind works firewall so chances of it picking it up something are quite slim. If there is a bit of software that can rule that out though ill try it!
Looking at getting a replacement mobo ordered to get here tomorrow so i can at least swap that in and start over. If the problem re-appears though with new board then im pretty stuck as its clearly not the motherboard then! These kind of errors are normally psu, mobo, memory or processor. As far as im concerned its not psu and memory (used 2 new psu's problem occured on both - memtest for 15hrs) so that leaves mobo or processor.
But the random-ness of the crashes tends for me to look at CPU as mobo will constantly be working whilst the cpu depends on the usage and crashes do only happen when its under some kind of load i.e. installing software, disk defrag, burning cd etc etc. But suppose you could say that when processor is under heavy load the mobo will be too and also questions asked as to why the stress tests didnt replicate problems it was only in windows normal running mode?!? grrr
>If there is a bit of software that can rule that out though ill try it!<
We have our own favourite tools, and Kaspersky is good. The next step would be ComboFix which does present a slight risk to your system. It would normally be used if HijackThis found nothing yet we knew you were quite heavily infected.
Recommend you attempt to get us a minidump or two, before the replacement mobo arrives.
If you decide to fit the mobo tomorrow, nothing lost in first running ComboFix. You could post back for all Combo details.
Ah, quick check! Is your RAM made up of two sticks? If yes, have you tried running on one of them to see if that resolves it? If no, re-fit the 2nd RAM, & retest. Any difference? i know, memtest reported it/them ok .. but it's a definite possibility!
Just re-logged on -- yes, both would be tested by Memtest.
However, ~no~ RAM Tester is perfect, & there are occasionally cases where running just one RAM stick will resolve an issue.
IMHO it's worth trying, as you appear not to have already done so. If no improvement, obviously try the 2nd stick.
Yes the CPU has the thermal compound, part of the oem heatsink and fan package that came with cpu.
Unless both new PSU's i have used have same fault then im doubting it is the psu to be honest.
Put one stick in last night and tried everything to fully test and it didnt BSOD. I did also read that some motherboards have a problem with certain memory when running in dual channel mode? and that PC's
Work fine with 1 in but when both in causes problems. Time will tell i suppose - but if that is the case then how do i verify if its the memory or the mobo causing the issues with dual channel?
would this make sense even though memtest didnt pick up errors? i mean does memtest test them whilst in dual channel mode or does it limit itself to testing each stick if you get what i mean.
Another quick question - could hard drive cause these problems? noticed today a slight clunking noise on startup before the drive spins into action, normal for sata drive? or possible indication that something not right? it passes all the tests ive run but still you never know.
really doubt this is hard drive related you would see more consistent faults if that were the case, you shoudl have yoru new mobo today, i owuld get that plugged in and tested to see if it resolves as everything else right now for me is just supposition
@ slam69 .. lol, just did a 'once in a blue moon' check of 'member rank', to see you passing me and disappearing over the horizon at a rate of knots! .. it's been good working alongside you ; )
Microsoft Windows XP is the sixth release of the NT series of operating systems, and was the first to be marketed in a variety of editions: XP Home and XP Professional, designed for business and power users. The advanced features in XP Professional are generally disabled in Home Edition, but are there and can be activated. There were two 64-bit editions, an embedded edition and a tablet edition.